REW Beta Release REW API beta releases

[moij]

This is data file.

 

[John Mulcahy]

Both measurements have significant time delays, so their phase responses go through many wraps, causing issues with the group delay calculation since that is the negative derivative of the phase. If you remove the time delay the group delay will be more sensible.
Edit: REW's group delay calculation should cope with the time delay, I'm looking into why it doesn't.

 

[John Mulcahy]

That group delay bug has been fixed for the next build.

 

[John Mulcahy]

Builds updated today (beta 102) with this change:

• Fixed: Group delay plots were incorrect when the IR peak was offset from t = 0 by more than a few ms

 

[stefano]

high John and thank you
i changed the old scarlett 2i2 for the motu mk5 and was taking measurements of the speakers.
when i test the eq section "measure with these filters" i do not obtain the predicted result. I remember that time ago i did always correctly applied with dsp device the eq filters and the test measurement before applying them was always consistent. in this moment i'm waiting for the new dsp device so i can't check the filters directly applied and i experiment only with rew, so i noted this problem. probably my problem, as i reinstalled rew, changed the soundcard and do not know what to look at. rew is at beta 102.

 

[moij]

I have more hard to explain measurement results

I made with DSP FIR filter to correct phase to 0 over full range but when I measure results phase is not corrected.
DSP was adding additional delay as expected, also magnitude was lowered in bass range as expected because lack of taps for FIR filter, so I am nearly sure DSP was applying FIR phase correcting filter. Also DSP in configuration shows phase correction filter was as expected, but I was not able to measure phase change with REW.
I used for phase correction FRI generation rePhase and REW measurement export where actual measurement was copied with "Min phase copy" function.

Where can be the error?
Is actually 8192 taps to few for this kind of phase correction or DSP have still not made FIR phase correction?

Here are before phase correction FIR and after measurements of magnitudes:


Here are before and after phases (thru "Min phase copy" what removes all delays)



Here is FIR phase correction impulse used on DSP:



Phase correction on rePhase what produced FIR impulse used on DSP:



Attached is also data file with original measurements and "Min phase copy" versions and FIR impulse.

 

[3ll3d00d]

does REW report anywhere where the estimated IR delay is after it is applied?

 

[John Mulcahy]

does REW report anywhere where the estimated IR delay is after it is applied?

The change will be reflected in the Cumulative IR shift figure on the Info panel and in the title of the t=0 adjustment dialog.

 

[John Mulcahy]

when i test the eq section "measure with these filters" i do not obtain the predicted result.

Which measurement did you select for Measure with these filters? It should be the one used when generating the filters. The resulting measurement will include a list of the filters used and the measurement they were obtained from in the notes.

 

[John Mulcahy]

I made with DSP FIR filter to correct phase to 0 over full range but when I measure results phase is not corrected.

This is the response of the filter you generated:

There's a lot of time delay in your measurements, which you could fix by adjusting the timing offset. I'm not sure what you intended or expected, but when you generate a minimum phase result the original phase of your measurement is discarded and a new phase response is generated from the magnitude. It is no longer the original measurement and it is not the same as removing time delays.

 

[moij]

So error is in my actions. How I get real phase of delayed measurements, is removing delay (Set t=0 at IR peak and update timing offset on Impulse) enough?

 

[stefano]

Which measurement did you select for Measure with these filters? It should be the one used when generating the filters. The resulting measurement will include a list of the filters used and the measurement they were obtained from in the notes.

Yes the resulting measurement reports the filters applied but the measurement is not the one expected. May this be caused by wrong soundcard settings? Could be an acoustical problem? The speaker is mounted on an open baffle. I tried a filter manual entered by me but is the same, the measured with filters graph is looking like the original one.
If the problem is mine it will be hard to be solved, not an expert.

 

[John Mulcahy]

May this be caused by wrong soundcard settings? Could be an acoustical problem?

Neither is likely to have any effect on your results. The measure with filters feature works as intended. Perhaps you could try connecting a loopback from the output to the input to eliminate any external factors and help isolate where the process is going wrong.

 

[stefano]

Neither is likely to have any effect on your results. The measure with filters feature works as intended. Perhaps you could try connecting a loopback from the output to the input to eliminate any external factors and help isolate where the process is going wrong.

Thanks for the attention, sorry.
I had no doubt that you and Rew were right. Changed only the speakers under test and all is right as before, perfect. I just a little separated the predicted and measured with filters graphs to compare them,..... identical. Now i must discover why that speaker is unresponsive to equalization. It was a nearfield measurement (0,5 cm)..... nothing else changed in the system......
Someone maybe has suggestions...

 

[sm52]

John Mulcahy, hello.​

More than 10 versions ago this phenomenon appeared. In Overlays, when hiding and opening some graphs through the left side list of measurements, before disappearing, the adjacent graph disappears for a moment, then appears. Then the one I want to hide disappears. If I want to hide the one on the left, it just disappears, and if the one on the right, what I described happens. In addition, the color of the graph that I clicked to hide goes to the one next to it on the left, then the color returns. But this is not always the case. It looks like when both are close to each other.

 

[sm52]

Second question. In windows 10 I was unable to receive data from the microphone if the driver is Java. REW has permission to access the microphone. If I switch to the ASIO driver, data from the microphone goes to REW, if I switch to Java, no data comes from the microphone. I tried many times, close, open Preferences window, close, open REW, disable microphone access, enable, disconnect sound card, connect to USB.

 

[John Mulcahy]

In Overlays, when hiding and opening some graphs through the left side list of measurements, before disappearing, the adjacent graph disappears for a moment, then appears.

If you remove a measurement using the selectors on the left the number of entries in the legend changes. That may change the size of the graph, if it does the graph will be redrawn.

 

[John Mulcahy]

In windows 10 I was unable to receive data from the microphone if the driver is Java.

That's not something REW controls. If you are using a WASAPI Exclusive Java entry there may be some other application that has taken exclusive access.

 

[sm52]

If you remove a measurement using the selectors on the left the number of entries in the legend changes. That may change the size of the graph, if it does the graph will be redrawn.

I don't delete a measurement, I click on its pad to the left of the graphs to hide or open to show its graph among other graphs. But, instead of just disappearing, there is a flashing. Maybe make another video to make it clearer?

 

[John Mulcahy]

By "remove" I mean deselect.